Mid-Range Forecasting of Solar-Wind: A Case Study of Building Regression Model with Space Weather Forecast Testbed (SWFT)

The Space Weather Forecast Testbed (SWFT) is developed by a team of space weather18 scientists and mathematicians at the University of Southern California (USC) and Jet19 Propulsion Laboratory (JPL) to foster the creation of models for space weather forecast20 by exploration of existing historic data using techniques of machine-learning. As an ex21 ample to demonstrate the potential power of SWFT, we present in this paper a multi22 linear regression based forecast model for solar wind. Solar wind is one of the key drivers23 for numerous physics-based models for space weather including thermosphere and iono24 sphere models. Many attempts have been made to produce forecasts for the solar wind.25 SWFT provides an unified framework for forecast model formulation, training and per26 formance assessment. In particular, the preparation of training and validation data by27 SWFT takes into account of realistic constraints on data latency and forecast lead time.28 In developing a solar wind forecast model, SWFT allows fast exploration of many meta29 parameters such as the list of predictive variables and their time history used in construct30 ing a model. We present the impact of meta-parameter selection, as well as, performance31 relative to existing solar wind forecast models.